Goal: The goal of clustering is to split the observations into groups such that the observations within each group are similar and the groups are different from one another.
In this type of clustering, we start with each observation as its own cluster (a leaf) and slowly combine them until we get to one cluster (a trunk). In the end, a dendrogram is used to help decide a good number of clusters to use.
Here is one example:
Moving up the tree, fuse similar leaves into branches. The more similar two leaves, the sooner their branches will fuse. The height of the first fusion between two cases’ branches measures the “distance” between them.
Allison Horst is the Artist in Residence at RStudio - how cool is that?! The following artwork is done by her and can be found on her github page.